Data Narratives
In the fall of 2024, I took a course titled Data Narratives. During the semester, I came across survey data collected by the Digital Society Project and the Varieties of Democracy Initiative. They interviewed “country experts”1 on the ways that netizens, governments, and political parties use the Internet and social media. For instance, they were asked questions about government control of social media, if the government uses it to spread misinformation to influence domestic and international populations for political reasons, the use of social media for political campaigns, etc.
1 A country expert is "typically a scholar or professional with deep knowledge of a country and of a particular political institution. Furthermore, the expert is usually a citizen or resident of the country. Multiple experts (usually 5 or more) code each variable."
Fig 1 An old notice tucked into a book titled Free Press, Free People in Brown University’s Rockefeller Library which sparked many projects over the semester.
Fig 2 January 6th Capitol Riots
My interest in the subject came from my ever-evolving and shifting relationship with the Internet and more recently, surveillance [FIG 1]. I was also interested in the increasing polarization and violent rhetoric on social media and the ways that it seemingly manifests offline (e.g. Incel2 and right-wing movements [FIG 2], the shift to the right in the 2024 election3 , etc.).
While I was not able to answer all of my questions with this dataset, it generated curiosity and questions for further research. Because the dataset is based off of the opinions of country experts rather than hard, quantifiable data (e.g. instances of hate crimes), the dataset, conclusions, and suggested relationships should be supported by other research.
Quantifying surveillance is extremely complicated with many limitations (e.g. methods of surveillance are generally undisclosed), but I don’t think this was wasted effort. The fact that experts were critical of their own countries suggests a healthy level of caution rather than total complacency.
2 Many of these online communities have extremely hateful rhetoric about women’s sexual and bodily autonomy, people of color, LGBTQ+ communities, etc.
3 Younger men voted decisively for Donald Trump which doesn’t surprise me as most of the top political commentators on major streaming platforms are right-leaning, and Trump made an effort in collaborating with streamers, even shouting them out after his victory. Source
Overall, which countries are more likely to control their citizens?
A broad overview of government control on digital media freedom provides context for political and netizen activity on a global scale. For instance, if a country is experiencing misinformation or divisive rhetoric, do netizens have enough digital media freedom to combat it?
In this context, “control methods” refer to censorship or blocking of information on the Internet, as well as the ability to shut down domestic access to the Internet. While some countries have the technical capacity to control its citizens, they don’t necessarily do these things in practice; the latter might be due to privacy laws6 that protect its citizens.
Click to enlarge.
According to expert opinion, countries with stronger privacy laws are less likely to control citizens’ access to the Internet. Perhaps these countries have privacy laws that also explicitly prevent the use of certain control methods, but it may also just reflect a culture of privacy. It’s difficult to know without looking at each individual country’s laws, but it’s reasonable to believe that countries with privacy laws also protect digital media freedom.
6 The question was “What does the legal framework to protect Internet users’ privacy and their data stipulate?” On a scale of 0-4, coders describe if the legal framework explicitly allows the government to access any type of personal data on the Internet or only in extraordinary circumstances.
Does government surveillance decrease instances of offline political violence?
Can an argument be made for the use of surveillance? One reason why government entities might use surveillance is to prevent offline violence. Experts believe that, over the last 23 years, government surveillance of social media, as well as the use of social media to organize offline violence, has increased over time.7
Click to enlarge.
7 Note that some major social media platforms were not developed until 2004–2006, such as Facebook, YouTube, and Twitter.
Most of the countries observed in the study saw an increase in government surveillance of social media content, but for the purpose of detailed analysis, I focused on a handful of countries that are of particular interest to the U.S. (and had enough experts for quality data).
Most of these countries did not see a huge increase except for Russia, which changed from “Somewhat Comprehensive” to “Extremely Comprehensive” levels of government surveillance. The most democratic countries stayed within the Limited/Comprehensive surveillance range whereas countries with authoritarian states like China, Russia, Vietnam, and Iran all moved towards the Extremely Comprehensive range.
It’s difficult to determine whether government surveillance of social media effectively prevents its use in organizing offline violence, especially given the unprecedented growth and widespread adoption of social media over the past 20 years. And, netizens may not know the extent of government surveillance in order to be deterred from using social media to organize violence.8 With that being said, the data above suggests that even the most authoritarian governments may not deter people from using social media to organize offline violence.
Social media is a double-edged sword. While social media can be a platform for free speech and self-expression, its users inadvertently give up a certain level of privacy. And, social media has become a destabilizing tool that can empower malevolent individuals and hate groups.
8 Even after Edward Snowden's leaks, the average American has carried on with their life.
Fig 3 Donald Trump infamously and falsely claims that Haitian immigrants are eating the pets of Springfield, Ohio residents during his debate with Kamala Harris.
Fig 4 A reminder of Trump's family separation policy which served to intimidate undocumented immigrants. As of November 2020, lawyers believe that they are unable to find parents for 666 kids. Source (via NBC News)
Does divisive political rhetoric manifest in offline violence?
Divisive rhetoric is another form of control and is really at the heart of my interest in surveillance. For this purpose, I specifically focused on the United States since my Internet feed mainly deals with American matters/discourse/users.
I couldn’t help but notice the harmful rhetoric throughout the course of the 2024 campaign cycle against multiple groups: immigrants [FIG 3–4], trans people, women9, pro-Palestinian protesters, etc. While this dataset certainly did not answer the question of why the election turned out the way it did (and Democrats will be scratching their heads for the next few years...), it was interesting to see what expert opinion was on the perceived frequency of political hate speech and offline violence, especially in the years leading up to the 2024 election.
Click to enlarge.
9 A new right-wing, manosphere slogan has emerged since Trump's 2024 victory: "Your body, my choice." Many respond, "so it's really not about being pro-life after all."
Fig 5 After the 2024 Election, Black students around the country received racist text messages. Source (via FBI)
Experts believe that political parties’ use of hate speech10 and the use of social media to organize offline violence has increased overtime11. This does not necessarily mean that instances of offline violence is increasing. In fact, violent crimes seems to be decreasing in the United States overall12, though, some datasets show conflicting or incomplete pictures).13
However, a troubling story emerges when we focus on hate crimes and politics. According to the FBI, hate crimes have increased over time.14 The Leadership Conference on Civil and Human Rights also claims that hate crimes have doubled since 2015, instances tend to increase during elections [FIG 5], and that white supremacists have become more active over the last four presidential elections (starting with Barack Obama). What’s even more troubling is that law enforcement agencies are not required to report hate crime data to the FBI, and 2022 saw the lowest level of participation.15 This means that we do not have the full scope of the problem.
So I have to wonder: what is leading to an increase in hate crimes? It’s hard to believe that it’s not related to the hate speech on TV and online which has fostered an “us versus them” mentality. I also believe that social media platforms have a major role to play since algorithms push users into echo chambers that encourage the use of divisive rhetoric. Social media has also made it easier for like-minded people to find each other and organize, for better or worse.
What about the government's role on the increase of hate crimes (or crimes organized via social media)? While offline violence is not limited to hate crimes, the Internet is prone to fostering extremist communities and hate speech which can manifest in offline violence. What good is social media surveillance if it is not decreasing offline violence? According to the second chart, it appears that surveillance does not deter people from using social media to organize offline violence. Are social media platforms not doing enough surveillance themselves? Do government entities have to hold these platforms responsible? Could the solution be something else entirely?
10 The question was “How often do major political parties use hate speech as part of their rhetoric?”
Hate speech is defined as “any speech that is intended to insult, offend, or intimidate members of specific groups, defined by race, religion, sexual orientation, national origin, disability, or similar trait.”
11 The question was “How often do people use social media to organize offline violence?”
12 Zamora, Karen, et. al. "Violent crime is dropping fast in the U.S. — even if Americans don't believe it," NPR, 12 February 2024. Source
13 Glider, Lucy. "What latest FBI data shows about violent crime," BBC, 18 October 2024. Source
14 "FBI Releases 2023 Hate Crime Statistics," U.S. Department of Justice, 25 September 2024. Source
15 "Cause for Concern 2024: The State of Hate," The Leadership Conference Education Fund, 13 May 2024. Source
Who do we hold responsible for all of this? What is the solution?
Both social media platforms and government entities are responsible for mitigating violence organized through social media. However, I also believe that we need to improve material conditions for the average person living in America. Many voters shifted to the right because they were struggling financially or otherwise believed that the status quo would not change under the Harris campaign.16 I believe that marginalized communities are used as scapegoats when others face poverty, lack of resources, or systemic discrimination. A reduction in the quality of education and health outcomes also makes it challenging for people to learn about new perspectives or challenging concepts, disempowering people from being able to talk about complex and nuanced issues. The United States' individualistic culture may push people to find community and group identity in sometimes negative spaces, only to push the blame on marginalized communities rather than hold the most powerful people, institutions, and cultural trends accountable for their suffering. In short, people are nuanced, and so are their reasons for voting and their understanding of the world around them. Without extending one another grace or believing in other people's ability to grow and change, we'll never be able to achieve the outcomes that this country deserves.
16 Many Americans felt that Kamala Harris did not differentiate herself enough from Biden.